On Fri, Oct 17, 2008 at 12:21 AM, Philippe Bourdeu d'Aguerre <[EMAIL PROTECTED]> wrote: > Le vendredi 17 octobre 2008, ottomeister a écrit : >> I wonder whether 'mkfifo' has become a shell builtin in the version of >> 'ksh' on this CentOS system. Philippe or Christian, perhaps you could >> run these commands in the CentOS 'ksh': >> >> rpm -q ksh >> builtin >> >> It would also be interesting to know what the exit status of 'mkfifo' is >> when it successfully creates a FIFO. > > mkfifo is not builtin and seems to return good status. The problem with mkfifo > in utconfig seems to be that the file already exists.
Yes, but it's not easy to see how that file can already exist. 'utconfig' creates it, uses it, and deletes it, all within a few lines of shell. The easiest explanation would be if 'mkfifo' was misbehaving but /usr/bin/mkfifo seems to be fine on machines here and in the invocations on your system. The fact that using a different 'ksh' binary makes the problem go away is what led me to wonder whether your 'ksh' was using a different implementation of 'mkfifo'. A builtin would be one way to do that, but apparently that's not what's happening. Another way of hitting this "FIFO already exists" problem would be to forcibly terminate 'utconfig' between the time that the FIFO gets created and the time it gets destroyed, but that's a very small window. And then, because the FIFO's name contains the ID of the 'utconfig' process, the run that fails would have to be using the same process ID as the earlier terminated instance. That's pretty unlikely. (Even so, 'utconfig' should defend against that possibility. This is a bug. It's just hard to believe that that this is what caused your 'utconfig' to fail.) Did you run 'utconfig' more than once? Did it hit this problem every time? What does 'ls /var/opt/SUNWut/tmp/utconfig.*' show on your system? Maybe there's a code path that I'm not seeing that leaves lots of these things lying around. > $ rpm -q ksh > ksh-20060214-1.7 I have ksh-20060214-1.4 on RHEL here. I'll see if I can find -1.7 and give that a try. OttoM. __ ottomeister Disclaimer: These are my opinion. I do not speak for my employer. _______________________________________________ SunRay-Users mailing list [email protected] http://www.filibeto.org/mailman/listinfo/sunray-users
